Originally published August 2021. Follow-up in November 2021.

Nouns.wtf (Twitter) dropped this week with an auction that sold for 613 E (a little under $2 million at the time of sale). A very impressive sale for a new generative NFT project - the best are lucky to mint their entire 10K collections for about that much.

The hype is in part due to its influencer packed ‘team’, but also its novel launch: one NFT auctioned every day, forever, with each sale funding a treasury that each NFT acts as an equal vote for. An interesting concept, but one seriously flawed when thought through.

Nouns is highly novel as a generative NFT project doubling as a DAO fund. It’s a pfpNFT project generated on-chain¹ from a randomized selection of base traits, but that’s where the similarities to other projects in the space (Cryptopunks, etc) stop:

• Rather than rarity tied to each trait’s random distribution, all the traits are randomized at an equal distribution. No given Noun is any more rare than another.

• Rather than a full drop, with the Noun’s traits unknown until purchase each Noun is fully visible when listed in an English auction, one minted every 24 hours; a new Noune is generated and goes live on auction immediately after the conclusion of the last auction.

• All the funds from the sale are released directly into a Nouns DAO, with each Noun NFT serving as one vote in the DAO. The DAO has no stated goals or mission statement, left completely open ended to be formed by governance vote.

On a closer review, the Nouns project is not really a generative NFT project at all, not as we’ve known them. It’s a self-seeding DAO, with a generative NFT tacked on as community identity (and branding). Very clever.

One could see it becoming a new standard for pfpNFT’s, making its use as an online identity with inalienable rights the primary factor, and not speculative aftermarket value. However, I have same grave doubts in the tokenomics design.

The Nouns are generated endlessly, one every day, each with an equal vote in the DAO. While the first few may auction into the millions, voting power is diluted daily, forever, which is a very strong deflationary pressure. One can also assume attention onto the project will fade as the DAO space becomes more saturated.

As both would lead to lower entry bids, it would be disadvantageous for the earliest investors. Buying in early is a higher risk for rights to control a smaller fund at a higher price per share, while guaranteed to have your share diluted in the long run. Smart money would wait and buy in late, after the fund is sizeable and proven, and price thoroughly deflated, achieving the same degree of power as someone who bought Noun 1.

As a friend put it, “Nouns is an experiment in punishing its earliest investors”

One could argue in the short-term the prices of the nouns will only go up over time, as the DAO will grow with every sale, thus making the relative value of buying into the DAO more lucrative, but this would not counteract its deflationary pressures. The earliest sales may see an upward term, but the long term must trend strongly downwards. Thus, one should expect to see the chart look like the infamous “Euthanasia Coaster”.

However, Nouns is organized as an open-ended DAO, meaning the community has the ability to recognize and amend these concerns by vote. I write this criticism both because I believe there is value in this concept that can be iterated on, and because the DAO could vote to save itself from its flaws.

• Fractional shares. It would go against the basic premise of Nouns as a democratic community, but vote & investment shares (eg proportional return on ventures) could be dilineated. A simple but inelegant solution would be proportional return based on initially buy-in cost, or more elaborately: a compounding discount rate applied on every new Noun’s serial number (e.g. Nouns are rated by age, the oldest noun has highest discount rate, newest have lowest). However either way would create a secondary market where Nouns are not equal but have different values attached them, which goes against the implicit equalizing vision of the project (though not sure this is necessarily something that needs to be kept).

• Sell-back. Set a certain price point where it’s deemed the Nouns have deflated to a significantly lower price floor. Allow older buyers the option to “sell-back” by returning their Noun in exchange for their initial cost of purchase. e.g. 7-day average sale more than 30% lower than it’s trailing 1-year average, any noun purchased more than 1-year eligible now for resale.

• Selling model fork. Perhaps a different selling model would be more elegant. A dutch auction (starts at a price point and trails downwards every X minutes until someone purchases)? This would have it unmoored from a 24-hour cycle, instead beginning the new auction the moment one is sold; the starting price point can be formalized as something like, 150% of the previous sale’s sale price, to prevent a whale sweeping the market.

• 23 hour cycle rather than 24 hours. This would allow the auction schedule to trail over time, fairly distributing any advantage for having an auction end during an individua’s sleeping timezone. Currently whatever disadvantage/advantage exists is baked permanently into the contract

• The Founder’s team hold executive veto rights during the ‘slow start’ of the project, to avoid early buyers with full control of the DAO from acting maliciously. This is a great idea, but it doesn’t seem there’s any formal definition of when it’s time to revoke control. It would be simple to decide on a date that it reasonably safe since the votes are diluted on a regular daily schedule, e.g. at 365 days.

I revisited the DAO in November. Besides the DAO failing to pivot its tokenomics and the Euthanasia Coaster chart playing out its process, they also demonstrated inability to pivot the DAO’s spending thesis, which was undermined by their high valuation.

This failure to pivot has resulted in repeat spending embarrassments, some of which have amounted to outright exploitation.

DAO’s are becoming more common in the generative NFT space, as the question of “where does all this money you’re making go?” comes into play; when the answer is ‘community development & marketing’, a democraticized DAO is an obvious trustless solution (though do I have my reservations²). These are DAO’s tacked onto the back of the NFT project, though; and most do not think of anything more creative than reinvestment in their own project’s marketing to drive secondary market value upwards, and some nominal investments to charities.

Nouns, however, is a DAO-first, and pfpNFT project second. There is no goal or mission statement, it’s entirely left open to the buyers in the community (though founder’s effectively maintain a consistent 10% vote of influence³).

If they continue to sell for the same prices for the rest of the year, they’ll be sitting with ~$350MM. What does a community do with all this? Invested wisely, it could be grown in a multi billion dollar fund. They could also act as whales dominating the NFT market.

They may lean back towards their pfpNFT roots though, and seek out developing ‘community projects’, like a videogame, or whatever. In all honesty, I find that less interesting than exerting themselves as a pure financial vehicle, with the Nouns acting merely as an avatar-identifier. We’ll see which direction of the scale they fall towards.

[1]: The psychedelic-pixel aesthetic looks like it should’ve come from Paper Rad, which reminds me what a shame it is they haven’t engaged the generative space beyond a low effort Foundation sale. Ben Jones characters, or a Jacob Ciocci on-chain gif art would be a natural fit - and the community they could produce in a neo-fort-thunder-discord would vibe shift the space forever.

[2]:

[3]: Founders are compensated with every tenth noun being minted Noun directly for themselves and distributed between the team in a multi-sig - effectively 10% of mints.